13

CONSUMER DEMAND IN NIGERIA1

by Ian G. Stewart
I. SOME GENERAL OBSERVATIONS
THEREare two good reasons fundamentally for inquiring into
consumers' behaviour. The fist of these is that data on con-
sumption comes nearer than data on either production or on
incomes to providing a yardstick or measure of current material
welfare. The second reason is that throughout a large part of
what people consume there is the strong stabilizing force of
inertia or habit or commitment at work which has encouraged
the economist to believe that prediction may be a permissible
exercise. In this paper I shall say something in Section I1 which
follows, about currently available estimates of consumer ex-
penditures, aggregate and partial, in the Federation of Nigeria,
and then proceed to discuss in Section 111 ways and means of
establishing in practice certain determinants of the level and
pattern of consumption in the next five to ten years.
It is a reflection of the difficulties which confront economists
and statisticians working in Nigeria that at the time of writing
(October 1960) there is still only one comprehensive account of
that country's national income. I refer, of course, to Prest and
Stewart's estimates for fiscal 1950-l.a The Federal Government
published in 1959 its first Economic Survey which contained
conjectural figures relating to the nation's aggregate product
and expenditure for the fiscal year 1956-7. The second sub-
stantive set of national income accounts to be published during
1962 has formed the basis of the paper contributed by
Dr. P. N. Okigbo in this volume. His estimates relate not
only to 1957, the benchmark period of his study, but also to
the years intervening since 1950-51, so that a continuous series
of a kind for aggregate output and expenditure is available.

partment of Agriculture, on whose bkhalf the stuir'deicrihed in Section 111 of

this paper is being conducted.
'A. R. Prest and I. G. Stewart, Tire Narional Income of Nigeria 1950-51,
H.M.S.O.. London Colonial Research Studies No. 1 1 .
~conomicSurvey of Nigeria 1959, Federal Government Printer, Lagos,
1959.
 308 AFRICAN STUDIES IN INCOME AND WEALTH
At a time when Nigeria is contemplating ambitious projects a
series of this sort can be invaluable, especially if it provides a
corrective to the too prevalent view that the Nigerian economy
has been growing rapidly in recent years. On the basis of the latest
agricultural information available, which relates to a triennial
period 1956-8, it would appear that gross farm output has
risen in constant 1950-1 prices by not less than 15 per cent but
not more than 42 per cent over the seven years 1950-1 to 1956-8.
When allowance is made for depreciation, which may be quite
small, and for population changes, which may have been quite
large, it is doubtful if net farm output per head can have been
rising at much over 12 per cent compound each year. Since
farm products constitute between 55 and 65 per cent of the
annual product, i.e. they exercise a predominant weight in any
index of production movements in Nigeria, to have achieved
an annual increment in G.N.P. as a whole of 4 per cent per
caput implies very remarkable progress indeed in the distribu-
tive, constructional and (embryonic) manufacturing sectors.
I emphasize this background to national income estimation
in Nigeria, f i s t of all because without it any discussion of con-
sumption would be out of perspective; secondly, in order to
draw attention to the comparative lack of material, which may
serve both as a word of caution to interpreters and of exhorta-
tion to practitioners in this important field of study.
There is one other general observation which may be helpful
in an understanding of Nigeria's present economic position in
relation to other nations in the African continent. Following
from the well-known propositions that as incomes rise (a) the
marginal propensity to consume and (b) the proportion of food
consumption expenditure to aggregate expenditure tend to fall,
a useful indication of relative stages in economic development
may be derived quite simply by comparing these two ratios for
different countries. One advantage of this formulation is, of
course, that like Lorenz curves of income distributions, inter-
national comparisons in ratio terms are less fallible than com-
parisons between values expressed in different currencies. In
theory, one could compare ratios of changes in private con-
sumers' expenditures to changes in disposable personal incomes
from country to country. In practice, however, it is still not
possible in most of the less-developed countries to calculate
marginal propensities; for one thing, the relationship between
 IAN G . STEWART 309
disposable personal income and gross domestic product is not
always sufficiently well established statistically. In Nigeria, all
that one can do is to look at the ratio of private consumption
expenditure to gross domestic product at factor cost: and make
a very crude comparison between the values of this ratio for
ditrerent countries. The proportion in the United Kingdom is
roughly two-thirds; in Tanganyika it has been in excess of
90 per cent, while Ghana and Nigeria consume annually and
privately about 85 per cent of their domestic product. According
to the Economic Survey of Nigeria 1959, the Federation devoted
roughly 5 per cent of all expenditure in 1956-7 to the Govern-
ment's current account, leaving gross investment at or near the
10 per cent level. The comparable figures in 1950-1 were 3 per
cent and 13 per cent respectively. This, however, cannot be
taken to imply that the rate of capital formation has been fall-
ing. In 1950-1 rather more than half of Nigeria's gross invest-
ment took the form of increases in sterling balances and other
assets held outside the country, while in 1956-7, with a reversal
in the previously favourable terms of trade, Nigeria was ex-
periencing a fall in sterling balances and a net inflow of longer-
term capital. It is perhaps a fairer comparison if domestic
investment alone is considered over the period, and here esti-
mates that indicate a doubling (from 6 to almost 12 per cent
of G.N.P.) are well documented: and consistent with the
writer's personal observation of the progress of a building boom
in Nigeria during the 1950s. While it is true that data of this
kind do not yet exist in su3cient quantity or quality upon which
to base any firm conclusions about aggregate savings or con-
sumption functions for Nigeria, it is di£Ecult to see at this stage
of the country's political and economic evolution how the ratio
of private consumption expenditure to total expenditure can be
expected to fall below 80-85 per cent in the next five to ten years.
11. PATTERN OF CONSUMER EXPENDITURE
Turning now to the pattern of private consumer expenditure
in Nigeria, one finds that the only period for which an analysis
' I t is quite possible that a 'correct* comparison of consumers' expenditure
at factor cost may be difficultfor reasons similar to those advanced by J. L.
Nicholson in his article 'National Income at Factor Cost or Market Price' in
The EcononficJournal, Vol. LXV, June 1955.
See Federation of Nigeria: Digest of Statistics, Vol. 7 , No. 1, and Vol. 8,
No. 3; also unpublished doctoral dissertation by A. Aboyade, Cambridge, 1960,
on 'Capital Formation in Nigeria'.
310 AFRICAN STUDIES I N INCOME AND WBAtTH

by commodity group is available is 1950-1, although the

Economic Survey of Nigeria 1959 estimated how much had been
spent in 1956-7 on imported consumer goods and how much on
home-produced goods and services, both figures being currently
pricedaggregat&.
Table I gives an indication of the proportions in which the
consumption expenditure of the private sector was divided
between differen; commodities in G50-1.
TABLE I
Expenditure by Nigeria

Privafe consumption expenditure

Food and drink
Clothing and footwear
Housing and water
Fuel and light
Other household goods
Cigarettes and tobacco
Travel (internal)
Education
Other services
Other goods
h e d forces income in kind
Total
Less Total imports (other than Government)
Total home-produced (other than Government)

It can readily be seen that out of total private expenditure

on goods and services a high proportion - over two-thirds -
went on food and drink, while the expenditure on consumer
durables, represented by 'other household goods' was very small
indeed - a mere 28 per cent. Despite the absence of later in-
formation cast in a comparable form, there is reason to believe
that this pattern of consumption has been altering, and in a
way that is consistent a priori with a positive rate of increase
in average income in various parts of the country.
One indicator is the changing level and composition of the
volume of goods imported into Nigeria since 1950. Expenditure
on imports of consumer goods at market prices appears to have
risen from about £86& million in 1950-1 to about £200 million
in 1957-8, which represents a rise in the proportion of con-
sumers' expenditure (at current prices) in respect of imports
 I A N G . STEWART 311
from about 15 per cent to nearly 25 per cent during the seven-
year period. In other words, any attempt at a general explana-
tion of consumers' behaviour in Nigeria must, if these figures
are reliable, give due weight to substitution effects and not
concentrate exclusively on income effects.
TABLE I1
Federation of Nigeria
Imports by classes, 1954-9
(E millions) (per cent)
Classes --- ---- -
1954 1955 1956 1957 1958 1959 1954 1959
0. Food 12.0 12.9 16.0 18.3 18.2 20.8
1. Beverages and
tobacco 4.4 5.0 5.3 5.5 5.6 5.8 160 15.7
2. Materials 1.5 1.7 1.9 1.9 2.0 2.0
--------
3. Fuels and lubri-
cants 5.6 6.5 7.3 8.2 8.9 10.4 5.0 5.8
4. +a1 and veg.
oils
5. Chemicals
6. Manufactures
7. Machinery and
transport equip-
ment
8. Misc. manufac-
tures
9. Parcel post, etc.
Total 1- ------
114.1 / 136.11 152.71 152.51 166.91 179.41 100 TI

Table I1 sheds some light on the way in which the pattern of

demand for imported goods in Nigeria has been altering over the
last five years. In spite of large increases in local production of
beer and cigarettes, the ratio of food and drink and tobacco im-
ports to the total imports scarcely changed from 1954 to 1959.
Imports of wheaten flour, refined sugar and meat have risen
during the same period by 80 per cent, 50 per cent and 280 per
cent respectively, while those of salt and gin have only risen by
9 per cent and 22 per cent respectively. The growth of bakeries
in many of the larger towns in Western Nigeria has been a
sigdcant step away from the older methods of 'food-process-
ing', an interesting development at a time when bread has
become, in the United Kingdom at any rate, an 'inferior' com-
modity with a negative income-elasticity of demand. Of the
larger volume of manufactured imports, textiles make up the
312 AFRICAN S T U D I E S IN INCOME AND W E A L T H
major consumer items, but during the period 1954-9 there has
only been a small increase in volumes of textiles imported.
There has, however, been a sixfold rise in the number of radio
receiving sets and nearly a threefold upsurge in the number of
motor-cars entering the country, and in the quantity of bicycle
parts for assembly.
The impression one gains is that imports of consumer goods
have certainly been keeping pace with the rate at which money
incomes have been rising, and have probably been rising more
rapidly due to net substitution effects (i.e. higher-grade food
and durable consumer goods more than compensating for
import replacement in the case of, say, tobacco for cigarettes
now wholly produced inside Nigeria) than because incomes
have, presumably, been rising.
At present, this is as far as one can proceed in the direction
of aggregate model-building in Nigeria. Once the national
income estimates extend from 1950-1 into a series for subse-
quent years, the prospects for research and study, and for pro-
jections of consumer demand in particular, should reveal a
range of most interesting problems.
If the current state of macro-economic data does not yet
provide scope for analytical exploration, there are happily
wider and increasing opportunities for making use of micro-
economic data, concerning the partial relationships between
household income, occupation and consumption expenditures,
classilied in the usual groupings.
The Federal Office of Statistics has carried out surveys of
urban wage-earning household expenditures in thirteen large
towns, starting with Lagos in 19534 and extending to each
of the three regions at intervals during the following five years.
Although the primary objective has been to obtain data on
which to construct retail price indices in principal urban areas,
the information that has been collected provides a basis for
estimating income elasticities of demand for diierent groups of
commodities. So far it would appear that there has been in-
sufficient time for the staff of the Federal Statistician's Depart-
'The first published survey dealt with wage-earning households in Lagos,
Ibadan and Enueu. Subseauent surveys have added Kano. ICaduna. Zaria. Warri.
Benin, Sapele, Iiorin, P o 6 Harcourt; Aha and Calabarto the list. e.g.'Urban
Consumer Surveys in Nigeria - Report on Enquiries into the Income and Ex-
penditure Patterns of Wage-Earner Households in Kaduna and Zaria 1955-56',
Lagos, 1959.
 I A N G . STEWART 313
ment, who are often hard-pressed by other commitments, to
conduct comprehensive econometric analyses based on the
family budget material. There is, however, one exception to
this statement. In a paper read before the 1958 Conference of
the Nigerian Institute of Social and Economic Research at
Ibadan, Mr. J. Heads, then one of the Federal Statisticians,
made the following observations:
'From Urban Consumer Surveys in a number of Nigerian
towns, it appears that food expenditure as a proportion of
total family expenditure on goods and services is almost
three-fifths for labourers, the lowest-paid occupational
group. For artisans, the proportion is less than one-half and
for clerks, the best-paid occupational group, it is only two-
fifths.' '
It should be added that 'best-paid' and 'lowest-paid' refer
merely to the three groups covered by these surveys, labourers,
artisans and clerks; it is not suggested that clerlts are better paid
than any other occupation in Nigeria. Heads used the sample
of 171 families interviewed in Ilorin, a predominantly Yoruba
town but situated just inside the Northern Region, to derive
partial correlation coefficients between (a) proportionate ex-
penditure on food and total expenditure and (b) proportionate
expenditure on food and size of family. From the partial re-
gressions he then obtained an estimating equation for the
income elasticity of demand for foodstuffs which, and I quote
again,
. . . 'Would suggest anincome elasticity of 0.65 at the average
Ilorin sample total expenditure level of 170.61 shillings (per
month). The Ilorin income elasticity of demand may not be
typical of the rest of Nigeria, but it can hardly be doubted
from statistical investigations of consumption patterns that
the overall income elasticity of demand for food is less than
unity here as in other countries. At higher income levels,
imported foods tend to become relatively more important
in food budgets and the income elasticity of demand for
home-produced foodstuffs is certainly even less than the
total income elasticity of demand for food.'
J. Heads, Urbanisation midEconomicProgress, in Nigerian Institute of Social
and Economic Research Conference Proceedings, December 1958 - published
by the Institute.
314 AFRICAN STUDIES IN INCOME AND WEALTH

As far as the writer is aware, this represents the first pub-

lished attempt to apply the apparatus of demand analysis with
which the work of Professor Richard Stone: 3. A. C. Brown,"
H. S. Houthakker and S. J. Prais have made us familiar, to
Nigerian data. Besides demonstrating what can be done for
urban areas with rather limited information, Heads's paper
suggests by implication two other directions in which 'scarce
statistical and economic skills' might be profitably employed.
The more important of these is the collectiou and analysis of
information culled from a small number of carefully selected
rural areas - only one in every five of Nigeria's 35 or 36 million
people live in towns having 5,000 or more persons - so that
important decisions of agricultural, fiscal and trading policy
may be based on at least a methodical, if not yet an accurate
assessment of the economic implications of the popular drift
to the towns. If it could be demonstrated statistically that, for
example, the marginal propensity at any given level of money
income of urban households to buy imported foodstuffs, tex-
tiles and other domestic articles for consumption is x per cent
greater than that of rural households at the same level of money
income, then important consequences for the balance of trade,
especially for the import of capital goods, would flow from any
marked tendency on the part of Nigerians to quit the farm and
seek employment in the many industrial estates that have be-
come increasingly a feature of the large towns in Western and
Eastern Nigeria, and to a lesser extent in the Northern Region.
While no organized body of systematic data is available on
consumption expenditures by rural households it is possible
to glean quite a lot of interesting and valuable information
from a disconnected series of widely scattered studies made
from time to time by individual researchers, notably anthro-
pologists. Recently, however, a more systematic approach to
problems of nutrition, by a medical field unit under the direc-
tion of Dr. B. M. Nicol: has shown that a comparative statis-
tical analysis of diets in widely separated and even remote
Richard Stone, The Meosurement of Consumers' Expenditure and Behaviour
in the United Kirzgdom, 1920-1938, Vol. 1, Cambridge, 1954.
J. A. C. Brown. 'The consumption of food m relation to household com-
position and income', Economelrico, 1954.
a H. S. Houthakker and S. 3. Prais, The Anolvsls of FamNy Budgets, Depart-
ment of Applied Economics Monograph No. 5, Cambridge, 1955.
&B.M. Nicol, 'The Calorie and Protein Requirements of Nigerian Peasants'
in British Journal of Ntrtritior~,Vol. 13, No. 3, 1959, pp. 293-320.
 IAN G. S T E W A R T 315
villages in Nigeria is well within the capacity of the Federal
Government's staff and budget. At the same time, the Federal
Statistics Office has been attempting to check some of the
figures in the Agricultural Census of 1955-8 by surveying con-
sumption in the villages selected for the Census. This is un-
doubtedly a desirable development and, provided national
accounting work is the main objective, one would like to
see a co-ordinated effort between economic statisticians and
nutritional experts so that a further series of sample villages
could be studied and fuller use made of the interesting results
so far obtained from several rural areas.
The other line of advance, suggested by the urban consumer
surveys of wage-earning households having incomes of less than
£350 per annum, is to study concurrently the pattern of con-
sumption expenditure of middle- or even upper-income house-
holds, say in the income range £350 to £1,500. Curiously
enough, at the same time (1959) that the Federal Office of
Statistics was planning to conduct just such a survey in Lagos
during 1960, a small team at the Department of Political
Economy in the University of Edinburgh was preparing to
undertake a similar but much less ambitious exercise in Ibadan
and subsequently in two Eastern towns?
111. SOME PRELIMINARY CONCLUSIONS
This section of the paper seeks to set out very briefly the
framework within which this academic enquiry into the re-
lationships between household income and expenditure is being
conducted, what the enquiry hopes to achieve and by what
methods it is proceeding. The fist thing to emphasize is that
the enquiry is restricted to investigating the demand in Nigeria
for agricultural products, whether domestically produced or
imported from overseas. On the demand side, therefore, atten-
tion is being focused on the consumption pattern of food, drink
and tobacco, and its relation to income and other variables such
as occupation and size/composition of family. One of the im-
portant aims in this has been to set an upper and a lower limit
' These surveys form p3rl of a nider study into the long-term dcvrlopmcnt of
supply and demand iur ngricullural products in Nlgcria being undertaken by
the Department in coniunction with the Niceriln Institt!tu of Econonlic Research.
Ibadai, for the United States Departmenfof Agriculture. I am indebted to ~ r :
Roderick Ogley, Senior Research Fellow at Edinburgh, for conducting the Pilot
Survey of Consumers' Expenditures on Foodstuffs in Ibadan, upon whose work
Section 111i ~ f r ais based.
316 AFRICAN STUDIES IN I N C O M E AND WEALTH

within which to approximate to estimates of future food re-

quirements in Nigeria, assuming different rates of increase in
per capita income over the next five or ten years. It may be
argued that estimates of consumption expenditure relating to
recent past experience set some kind of lower limit, save in the
event of a drastic deterioration in the terms of trade. If the
relation between export and import prices worsens slowly - as
it has tended to do on balance over the last five years - the
present level of consumption expenditure can still be main-
tained by increments in volumes and for a number of years by
drawing upon the fairly comfortable reserves of sterling, some-
where in the region of £200 millions at the present time. The
really interesting question is posed by the upper limit to certain
kinds of consumer spending over the coming decade, since a
high marginal propensity to import consumer goods could have
serious implications for Nigeria's rate of capital accumulation.
One way of setting an upper limit to this prediction of future
consumption expenditures is to assemble data relating to those
households which at the present time have a high level of con-
sumption including a propensity to consume imported goods
that is well above the national average. A pilot survey along
these lines has now been made, duing the second quarter of
1960, in Ibadan, in which some 160 households with salaries
ranging from £300 to £2,500 per annum provided information
on household composition, income and a detailed list of daily
expenditure on foodstuffs over a period of two weeks. Con-
sidering that Nigerians, and all but 6Fteen or twenty of the
households approached were Africans, are every bit as conscious
of the considerations that move the well-to-do to keep silence
as are the sophisticates of any other economy, a rate of response
of 70 per cent can be described as encouraging. Simultaneously
the Federal Government Statistician has been conducting a
broader and more extensive survey of middle-income households
in Lagos, embracing the whole range of consumption expendi-
tures and malting observations in each of a large number of
households for one month out of every three, so that seasonal
variations can be smoothed.
While it is still too soon to say anything about the official
survey of middle-income households and premature to go into
any detailed evaluation of the results so far analysed from the
Ibadan Pilot survey during March-June, 1960, it may be of in-
 I A N G. STEWART 317
terest to note one or two tentative conclusions that this smaller-
scale enquiry has thrown up, conclusions which have been care-
fully checked against the results of a comparable survey carried
out in two towns in the Eastern Region of Nigeria early in 1961.
(a) The evidence so far collected lends some support to the
view that traditional Engel curves can be fitted to expenditures
on food by Nigerian middle and upper-income households.
Although this evidence is subject to a wide margin of error, it
does appear from the general co&guration in Chart I below
that the demand for foodstuffs is income-elastic at lower ranges
of income, but tends towards an inelastic relationship as income
per consumption-unit (i.e. adult equivalent) rises above £360.
This would imply a household average income in the range
£75&£1,000, depending on the estimation of 'adult-equi-
valence'. The really important point to emphasize is not the
exact measurement of the range of income but the impression
that techniques being developed for the analysis of demand in
'Western' economies can with suitable modification be applied
to Nigerian problems of this kind.
(f 1 per month)
?
L

z.
7
incovvle per ernsumption- unit

(b) The proportion of household expenditure on food that is

made up of imported foodstuffs, drink and tobacco appears to
be 12-15 per cent over the lower range of salaried households
318 AFRICAN STUDIES IN INCOME AND WEALTH
in Ibadan, i.e. those with incomes from £300 to £700. There-
after there is a sudden upwards adjustment to around 30 per
cent over the income range £75&&1,500, and then a slight
decline in the ratio of imported to locally produced foodstuffs,
etc., consumed by upper-income households.
(c) While it would be extremely unwise to generalize at all
on the basis of one small survey conducted in Ibadan, one of
the explanatory variables to which more attention might pro-
fitably be directed in future work of this kind is that of the
household size. In this survey household size seems to vary
inversely with income even after due allowance for under-
recording of total receipts.
These are but a few preliminary and tentative conclusions
about the kind of hypothesis one might reasonably try out in
the next stages of analysing consumers' behaviour, and it is
expected that during 196163 some of these ideas can be com-
pared with the results of another survey which the University of
Edinburgh is undertaking in Ghana.
The next stage will be to assign 'weights' to the different
income groups, occupational groupings and demographic cate-
gories, i.e: to attempt to move from the particular survey results
in the duection of assessing general trends in the macro-
economic relationship between aggregate income and expendi-
ture on foodstuffs, etc. There is plainly a very long way to go
before one can approximate to the degree of sophistication
attained by demand analysis in more-developed countries, and
the crying need is for suitable data, and therefore for a vigorous
policy in support of the statistical work ably begun and carried
on in spite of immense difficulties by the statistical services,
federal and regional.